Determination of Allyl Groups in Polyallyl Ethers and Esters H.4ROLD .\I. HOI-D
~ Y D J.
ROBERT ROACH, General .lfil/s, Znc., Vlinrceupolis, .\firin.
:\llyI ether and allyl ester groups undergo a quantitative reaction with various reagents for the addition of halogen. I n this study it has been found that allyl acetate, allyl phthalate, and triallyl glycerol can be accurately analyzed for allyl groups by numerous methods-for example, by the Wijs, rapid Wijs,
I
Rosenmund and Kuhnhenn, and bromine method*. The Kaufmann method gave results which were approximately 10% low. The number of allyl groupsthe degree of substitution-in polyallyl derivatives can be determined by the same methods. The most rapid and simplest is the rapid Wijs method.
S ETHERIFICATIOS and esterification of polyhydric
ses were first carried out on known compounds. For this study highly purified samples of allyl acetate, allyl phthalate (the diester), and triallyl glycerol were prepared. The results of a study of five different analytical methods on these three compounds are tabulated in Table I. The analyses of the allyl derivatives should be performed as soon as the substances have been prepared, although inhibitors such as hydroquinone can be added in order to decrease polymerization. Storing in the cold under an inert atmosphere such as nitrogen or carbon dioxide likewise decreases polymerization. Samples of triallyl glycerol showed a decrease of 2 to 4 units in iodine values (three different analytical methods) upon standing for 2 days a t room temperature in the absence of inhibitors. Skell and Radlove (9) observed that methyl ricinoleate and methyl ricinelaidate gave theoretical iodine values by the Wijs method but gave high values by the rapid Wijs method, whereas methyl o-propionylricinoleate behaved normally. They concluded that this anomalous effect is due to the presence of the free hydroxyl group. Consequently i t was of considerable interest to note that a sample of or,d-diallyl glycerol, prepared from glycerol dichlorohydrin and allyl alcohol, gave iodine values of 291.3 and 292.0 by the rapid Wijs method as compared to a theoretical value of 295.1. Likewise it can he observed from the results in Table I1 that the presence of free hydroxyl groups in partially allylated allyl sucrose and allyl starch offers no difficulties in analysis.
alcohols, the degree of reaction or the degree of substitution may be ascertained by a Zeisel determination for the ether groups or a saponification equivalent for the esters. With the increasing interest in polyallyl ethers and polyallyl esters, a more convenient method of analysis for degree of substitution is highly desira ble. The purpose of this work was t o detemine the applicability of various known methods for the determination of unsaturation to the analysis of allyl ethers and allyl esters, and to establish a reliable method for the determination of the degree of suhstitution of polyallyl derivatives. CHE3IICA LS
-illy1 acetate was prepared and purified according to known methods; boiling point 105O, n',"" = 1.4015. -Illy1 phthalate ivas prepared and purifiedoaccording to known methods; boiling point 126" a t 0.7 mm., nZ,J = 1.5169. Allyl sucrose was prepared according to the procedure of *Sicholsand Yanovsky ( 7 ) ,with slight modifications. hllyl starch was prepared by the procedure of Sichols arid cowith a few modifications. n-orkers (67, Triallyl glycerol was prepazed according to the method of Sichols and Yanovsky ( 7 ) ; n'," = 1.4501. REAGENTS AND GENERAL PROCEDURE
In all cases approximately 0.1-gram samples mere used, and with the relatively volatile allyl acetate it was necessary to have the reaction flasks containing 10 ml. of chloroform cooled in a refrigerator in order to eliminate loss of sample. Rapid Wijs. The usual Wijs reagents were prepared while the mercuric acetate catalyst was prepared and the analyses w x e carried out according to the procedure of Hoffman and Green
DISCUSSION
From the iodine values obtained on pure known compounds it is apparent that numerous methods can be used for determining the number of allyl groups in a polyallyl ether or ester. Of these, the simplest and most rapid (4 minutes) is the rapid Wijs method, which in all instances gave iodine values less than 1%lower than the theoretical value. The ordinary Wijs method gave slightly lower values, but this is not surprising since McCutcheon (6) found that the iodine values of ethyl linoleate and methyl linolenate, obtained by the Wijs method, were 162.4 and 257.3 as compared to the theoretical values of 164.7 and 260.5.
(3).
One-Hour Wijs. The reagents were prepared and the analyses were carried out as outlined in "Official and Tentative .\lethods of Analysis'' ( 1 ) . Kaufmann. ?n approsiniately 0.1 -Y bromine solution i n methanol which had been saturated with sodium bromide \vas prepared and the analyses were made according to the directions of Kaufmann and Hartwig 141. Rosenmund and Kuhnhenn. The pyridine sulfate dibromide solution was prepared and the analyses were made according to the procedure of Rosenmund and Kuhnhenn (8). Bromine Method. A 0.5 *V solution of bromine in chloroform (10 ml.) was added to a cold flask containing a 0.1-gram sample dissolved in 10 ml. of chloroform. The flasks were placed in a refrigerator at approximately 4' and after approsimately 10 minutes the stoppers were sealed with a few milliliters of potassium iodide solution and the flasks were allowed to stand a t 4 " for the desired time. If the sealing is carried out before the flasks and contents have completely cooled, the potassium iodide solution will he drawn into the reaction flask.
'Table I. Comparison of Iodine Values Obtained on Known Compounds AI et h od
EXPERIMENTAL RESULTS Kaufmann
I n order to determine the validity of the various analytical procedures for determination of unsaturated groups, the analy-
Theoretical iodine number
158
Allyl Phthalate
Allyl Acetate
Triallyl Glycerol
207.5,209.6 207.8,208.4 204.0,205,0 205.5,204.0 204.0,203.0 208 0,207.0 181.1.182.8 206.1
252.1,251.2 251.7,251.8 249.7,251. O 252.3.252.0 249.2,248.3 251 .O,249.0 225.3.227.4 253.8
355:4;365.7
352:0;352. 6 354: i;354 .o 319.6.326.9 358.3
159
V O L U M E 19, NO. 3, M A R C H 1 9 4 7 Table 11. Analysis of Allyl Sucrose and Allyl Starch Method 1-hour bromine method 2-hour bromine method 3-hour bromine method Rapid Wijs 1-hour Wijs Rosenmund a n d Kuhnhenn Iiaufmann
Allyl Sucrose Iodine KO. D.S. 268.3,268.1 6.25 270.6,269.6 6.32 271.0,272.2 6 . 4 2 273.0,274.0 6.45 272.1,271.3 6.42 269.1,264.3 6.30 6.10 231.3,233.8 4.95
Allyl Starch Iodine KO. D.S. 183.8,181.7 1.62 180.6,181.0 1.60 180.4,180.5 1 . 6 9 176.7,177.4 1.55 175.8,176.6 1.54 174.6,175.6 1.53 143.5,152.5 1.23
It is of interest that the Rosenmund and Kuhnhenn method gave results which differed from the theoretical values by approximately 1%, since Earle and Milner ( 2 ) found that this method gave results appreciably low for all vegetable oils having iodine numbers greater than 100. In all instances the Kaufmann method gave results approximately 10% lower than theoretical or the average of the other methods. The degree of substitution for triallyl glycerol as calculated from the iodine values varied between 2.90 and 2.96, with the rapid Wijs method giving the degree of substitution of 2.96.
In applying these methods to the determination of the degree of substitution of allyl starch and allyl sucrose the iodine values were in good agreement, as were the degree of substitution values. LITERATURE CITED (1) hssoc. Official Agr. Chem., Official and Tentative Methods of Analysis, p. 90 (1940). (2) Earle, T. R., and Milner, R. T., Oil & Soap, 16,69 (1939). (3) Hoffman,H. D., and Green, C. E.,I b i d . , 16,236 (1939). (4) Kaufmann, H. P., and Hartwig, L., Ber., 70B,2554 (1937). ( 5 ) McCutcheon, J. W., IND.ENG.CHEM.,AKAL.ED.,12, 465 (1940). (6) Xichols, P. L., Jr., Hamilton, R. >I., Smith, L. T., and Yanovsky, E., l n d . Eng. Chem., 37, 201 (1945). (7) KJichols,P. L., Jr., and Yanovsky, E., J . Am. Chem. SOC.,67, 46 (1945). (8) Rosenmund, K. W.,and Kuhnhenn, TT., Z . C'ntersuch. S a h r . u . Genussm., 46,154 (1923) ; Analyst, 1924,105. (9) Skell, P. S., and Radlove, S. B., IND. EJG. CHEM.,ANAL.ED., 18, 67 (1946). PRESESTED before the Division of Analytical and Llicro Chemistry a t the 110th Meeting of the A 3 f E R I c A x CHEMICAL SOCIETY, Chicago, Ill. Paper 73, Journal Series, General Mills, Inc., Research Laboratories.
Polarographic Determination of Vanadium in Steel and Other Ferroalloys JAMES J. LINGANE AND LOUIS MEITES, JR. Department of Chemistry, Hurvard University, Cambridge 38, Muss.
A method is described for the polarographic determination of vanadium in steel and other ferroalloys, based on the removal of interfering elements by electrolysis with a mercury cathode from a dilute sulfuric-phosphoric acid solution, and subsequent measurement of the anodic diffusion current produced by the oxidation of C4 to 4-5 vanadium in a supporting electrolyte containing 0.5 to 3 N sodium hydroxide and 0.1 M sodium sulfite. Results obtained in the analysis of ten Bureau of Standards samples of ferroalloys, containing from 0.01 to 2% vanadium and large amounts of manganese, molybdenum, tungsten, chromium, nickel, copper, and other common alloying elements, were in excellent agreement with the bureau's values.
T
HE polarographic determination of vanadium in steel and
determination of vanadium in steel is that of Stackelberg
other ferroalloys described in this paper is based on the measurement of the anodic wave produced by the oxidation of f 4 to $ 5 vanadium in a strongly alkaline supporting electrolyte. The general characteristics of this wave have been described in a previous paper (6). Iron and other interfering elements are removed by electrolyzing a solution of the sample in a phosphoricsulfuric acid solution with a mercury cathode. Of the elements likely to be present in ferroalloys this procedure leaves in the solution, in addition to vanadium in the + 3 state, only molybdenum, tungsten, titanium, uranium, columbium, tantalum, aluminum, and a trace of manganese. Kone of these elements interferes in the determination of the vanadium. The residual solution is treated with hydrogen peroxide and then with sulfite (sulfurous acid) in excess to convert the vanadium to the +4 state, and is made up to a known volume. An aliquot portion of this solution is added to a known volume of air-free 1 N sodium hydroxide in a polarographic cell and the polarogram is recorded. 4 complete determination requires only about 90 minutes' elapsed time. Results obtained with a variety of Bureau of Standards samples demonstrate that the method is as accurate as the classical methods for steels containing a few per cent of vanadium, and with very small amounts of vanadium the polarographic method is probably more reliable. The only previously reported method for the polarographic
et al. (11). These authors used the classical sodium hydroxide
precipitation method for separating iron and various other elements from +?Ivanadium, and the latter was finally determined in an ammoniacal supporting electrolyte. This method is rapid, but it involves the danger of coprecipitation of some of the vanadium, particularly when much manganese is present, and the results quoted by Stackelberg et al. are less accurate than those obtained by the present method. I n preliminary work the electrolytic separation with the mercury cathode was carried out by the usual technique from a dilute sulfuric acid solution (1, 9), but difficulties arose when much manganese and molybdenum were present. Manganese was partially oxidized to manganese dioxide and permanganate ion a t the anode, and molybdenum tended to precipitate (probably hydrous hloO2) during the electrolysis. By adding phosphoric acid as recommended by Chlopin (2) these difficulties were avoided. EXPERIMENTAL TECHNIQUE
The cell used for the electrolytic separation was of the Melaven type (1,9,10); it had a capacity of about 200 cc. and the area of the cathode mercury was about 25 sq. cm. A coil of platinum wire served as anode. Stirring was accomplished by a glass propeller stirrer whose blades were partly immersed in the mercury, so that the mercury-solution interface was well stirred. A current of 3 to 5 amperes was used, a t a total applied e.m.f. of about
